[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15439619#comment-15439619 ] Lars Hofhansl commented on PHOENIX-258: --- Chatted with [~giacomotaylor], and [~mujtabachohan]. Turns

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15412436#comment-15412436 ] Lars Hofhansl commented on PHOENIX-258: --- That does not at all match my experience. It only needs to

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15388898#comment-15388898 ] James Taylor commented on PHOENIX-258: -- [~lhofhansl] - I would have expected a bigger gain for only

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15388655#comment-15388655 ] Mujtaba Chohan commented on PHOENIX-258: [~lhofhansl] Tested on a table with 300M rows 230GB

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15325662#comment-15325662 ] Hudson commented on PHOENIX-258: FAILURE: Integrated in Phoenix-master #1254 (See

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15314380#comment-15314380 ] Lars Hofhansl commented on PHOENIX-258: --- Actually it's all good. master and -1.0 are back to

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15314373#comment-15314373 ] Lars Hofhansl commented on PHOENIX-258: --- Deleted the two branches. Sorry about this. I think I had

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15313260#comment-15313260 ] Lars Hofhansl commented on PHOENIX-258: --- There are some more tests failing, all test fixes I think

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15313163#comment-15313163 ] Hudson commented on PHOENIX-258: SUCCESS: Integrated in Phoenix-master #1242 (See

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15312995#comment-15312995 ] Josh Elser commented on PHOENIX-258: bq. I might have inadvertently created a 4.0 and 4.x-HBase-1.x

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15312985#comment-15312985 ] Lars Hofhansl commented on PHOENIX-258: --- Done. I might have inadvertently created a 4.0 and

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15312975#comment-15312975 ] Lars Hofhansl commented on PHOENIX-258: --- There are some test failures. Fix soon. > Use skip scan

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15312893#comment-15312893 ] Hudson commented on PHOENIX-258: FAILURE: Integrated in Phoenix-master #1241 (See

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15311600#comment-15311600 ] James Taylor commented on PHOENIX-258: -- +1. Awesome job, [~lhofhansl]. I love the trick you thought

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15311543#comment-15311543 ] Lars Hofhansl commented on PHOENIX-258: --- bq. I think what we need is an int[] indexed by the slot

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15311539#comment-15311539 ] Lars Hofhansl commented on PHOENIX-258: --- BTW... Here's the HBase logic in FilterList to AND filters

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15311408#comment-15311408 ] Lars Hofhansl commented on PHOENIX-258: --- Actually, so here's a case that does not work: 1,

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15311293#comment-15311293 ] Lars Hofhansl commented on PHOENIX-258: --- bq. For example, if we have a schema like VARCHAR, INT,

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15311262#comment-15311262 ] Lars Hofhansl commented on PHOENIX-258: --- bq. What if you have these three rows (assume INT, 2 byte

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15310723#comment-15310723 ] James Taylor commented on PHOENIX-258: -- I think what we need is an int[] indexed by the slot

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15310668#comment-15310668 ] James Taylor commented on PHOENIX-258: -- What if you have these three rows (assume INT, 2 byte

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15310651#comment-15310651 ] Lars Hofhansl commented on PHOENIX-258: --- bq. The reverse scan case is tricky. Padding with a single

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15310630#comment-15310630 ] Lars Hofhansl commented on PHOENIX-258: --- "cannot" should "can only"! > Use skip scan when SELECT

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15310640#comment-15310640 ] James Taylor commented on PHOENIX-258: -- (and yes, you're correct - don't need to worry about

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15310636#comment-15310636 ] James Taylor commented on PHOENIX-258: -- Yes, good point, [~lhofhansl]. The reverse scan case is

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15310627#comment-15310627 ] Lars Hofhansl commented on PHOENIX-258: --- So specifically: bq. You have to pad with some number of

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15310571#comment-15310571 ] Lars Hofhansl commented on PHOENIX-258: --- [~giacomotaylor], VARBINARY can only be last in a PK,

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15309418#comment-15309418 ] Lars Hofhansl commented on PHOENIX-258: --- Actually before we pull the trigger I want to add an

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15309327#comment-15309327 ] Lars Hofhansl commented on PHOENIX-258: --- Nope. The next part will start with something >= 0x00. So

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15309316#comment-15309316 ] Lars Hofhansl commented on PHOENIX-258: --- Turns out there's a problem for fixed length fields when

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15309272#comment-15309272 ] Lars Hofhansl commented on PHOENIX-258: --- Hmm... isOrderPreserving is calculated like this:

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15309256#comment-15309256 ] Lars Hofhansl commented on PHOENIX-258: --- Lemme do those two. Was thinking on the

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15308792#comment-15308792 ] James Taylor commented on PHOENIX-258: -- +1. Looks very good, [~lhofhansl]. Couple of optional items,

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15308212#comment-15308212 ] Lars Hofhansl commented on PHOENIX-258: --- Makes sense. Thanks for the case where it would not work.

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15308159#comment-15308159 ] James Taylor commented on PHOENIX-258: -- Thanks, [~lhofhansl]. Please file a separate bug for the RVC

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15307059#comment-15307059 ] Lars Hofhansl commented on PHOENIX-258: --- Thanks for the full description! I'll add that code later

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15306800#comment-15306800 ] James Taylor commented on PHOENIX-258: -- This is looking very good, [~lhofhansl]. A couple of minor

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15306299#comment-15306299 ] Lars Hofhansl commented on PHOENIX-258: --- Wow... Reverse scans are slow. The speedup for a reverse

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15306290#comment-15306290 ] Lars Hofhansl commented on PHOENIX-258: --- Oh I see what you mean. Yeah the optimization would not

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15305772#comment-15305772 ] Lars Hofhansl commented on PHOENIX-258: --- Thanks [~giacomotaylor] -v7 has the follwing: * Renamed

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15305654#comment-15305654 ] James Taylor commented on PHOENIX-258: -- This is awesome, [~lhofhansl]! Thanks so much for working

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15304850#comment-15304850 ] Lars Hofhansl commented on PHOENIX-258: --- [~giacomotaylor] where/how is COUNT(DISTINCT ...)

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15303122#comment-15303122 ] Lars Hofhansl commented on PHOENIX-258: --- * g1, g2, g3 do not even have to be ordered in key order *

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15301119#comment-15301119 ] Nick Dimiduk commented on PHOENIX-258: -- In my particular case, itemizing the leading column also

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15300729#comment-15300729 ] Lars Hofhansl commented on PHOENIX-258: --- [~ndimiduk], FYI. You mentioned you had some stats. Wanna

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15300657#comment-15300657 ] Lars Hofhansl commented on PHOENIX-258: --- I'll attempt a combined patch. Should be just using a

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15261389#comment-15261389 ] James Taylor commented on PHOENIX-258: -- [~lhofhansl] - had any spare cycles to work on this and if

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15242195#comment-15242195 ] Lars Hofhansl commented on PHOENIX-258: --- It's a little more complex than that. We need to increment

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15241851#comment-15241851 ] James Taylor commented on PHOENIX-258: -- I don't think we need another filter for this optimization -

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15241775#comment-15241775 ] Lars Hofhansl commented on PHOENIX-258: --- If this means nobody else will be looking at it, it's

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15241760#comment-15241760 ] James Taylor commented on PHOENIX-258: -- [~lhofhansl] - you ok if we assign this to you (as you've

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15212158#comment-15212158 ] James Taylor commented on PHOENIX-258: -- See PHOENIX-2797 for a good description and rationale. >

[jira] [Commented] (PHOENIX-258) Use skip scan when SELECT DISTINCT on leading row key column(s)

[ https://issues.apache.org/jira/browse/PHOENIX-258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14136557#comment-14136557 ] James Taylor commented on PHOENIX-258: -- [~kbuzsaki] - this would be a good one for